This paper was prepared for presentation at the SPE/ICoTA Coiled Tubing & Well Intervention Conference & Exhibition held in Houston, Texas, USA, 2223 March
2016.
This paper was selected for presentation by an SPE program committee following review of information contained in an abstract submitted by the author(s). Contents
of the paper have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material does not necessarily reflect
any position of the Society of Petroleum Engineers, its officers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the written
consent of the Society of Petroleum Engineers is prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300 words; illustrations may
not be copied. The abstract must contain conspicuous acknowledgment of SPE copyright.
Abstract
The causes and associated remediation practices of most coiled tubing failures are well-known and
detailed in multiple technical papers (Crabtree 2008, van Ardrichem & Larsen 2002, Burgos et al. 2007).
One type of failure that has more limited exposure is excessive wall loss due to abrasion, occurring
primarily in chrome completions. Significant wall loss over a large area typically requires the string be
retired from service. On occasion, this damage has occurred during only a few operating runs. Retiring
strings early is financially unacceptable, and in certain global locations, replacement strings can take
months to arrive, resulting in a failure to meet operational objectives. This paper will detail the
background on this type of failure observed in Azerbaijan, New Zealand and the USA, the failure analysis,
results from pipe inspections, mitigation plans, and their associated results in reducing the frequency of
incidents.
Analysis of incidents in the three markets started with an early notice of wear on a concentric string
for offshore use in chrome completions. This wear detection resulted in a full string inspection to
determine future life. While this string was retired, mitigation practices included mandating a wetted
and/or lubricated wellbore and greater tracking of wall loss. The second market experienced a significant
number of cases of premature failure, primarily on 1 -in. coiled tubing (CT) run in chrome completions.
Significant tracking of pipe movement versus location of the damage was completed to optimally target
the use of friction-reducing lubricants. The third market used a 2-in. concentric string for 14 runs in a
single L80 well. Here, flat spots exceeded 60% of wall loss, and several compressive failures occurred on
the reel. In addition, galling with material transfer was observed. Mitigation practices include the use of
pipe straighteners, wetted and lubricated completion in areas of high contact force.
Mitigation practices have significantly reduced the extreme wall loss issues in the first market, and
similar programs have been recently implemented in the other two. Features common to all wells, such as
deviation profile, will be presented to identify high-risk wells and permit pre-job mitigation planning.
This paper covers data from 19 wells and various mitigation practices and the associated positive results.
The information will enable operators to better plan for success with CT well intervention
operations.
2 SPE-179096-MS
Introduction
Coiled tubing failures associated with severe wall thickness reduction (not including corrosion) have been
observed to be associated with two mechanisms:
Fine particles erosion, associated with the erosive effect of sand used for fracturing jobs shown in
Fig. 1.
External abrasion due to friction between the coiled tubing string and the well completion (Fig. 2).
Failures by external abrasion are characterized by the following:
Severe wall thickness reduction normally localized in a short length (approximately 3 to 60 inches,
Fig. 3).
High external roughness and galling on the abraded surface (Fig. 4). Galling is a severe form of
wear on sliding components characterized by localized macroscopic material transfer, i.e., frag-
ments or surface protrusions that are visible on either or both surfaces.
SPE-179096-MS 3
Failures at localized bending points within the string; for example, areas close to bias welds (Fig.
5a), or pipe waviness (Fig. 5b) or close to the bottom hole assembly (BHA).
Figure 5a & bAbrasion Close to Bias Weld (left) & General Pipe Waviness (right)
Incident History
On review of a failure incident database from 2005 (date of first noted abrasion incident) to 2015, only
17 from 354 failures (5%) were attributed to external abrasion. Five of the failures occurred within a short
distance, less than 15 ft, from the BHA. These failures likely experienced abrasion throughout all well
movements due to excess residual curvature. Of the 12 failures that occurred away from the BHA, 9 cases
(75%) were in chrome completion. This paper will also discuss two other cases (both in Azerbaijan) where
initial light wall gouging was noted as the onset of abrasion and the strings were prematurely retired.
Fig. 6 shows the distribution of external abrasion failures by region and country. External abrasion was
more common on regions where the well completions were mainly 13% Cr (11 out of 17 total or 65%).
4 SPE-179096-MS
Within the authors organization, failures associated with external abrasion are not common, as shown
in Fig. 7. The number of cases per year in the period 2005 and 2015 was between 1 and 3, except for 2014
that presented 4 failures associated with external abrasion.
Despite the low number of failures associated with external abrasion, incidents have occurred with very
low running footage and occurred very quickly, requiring additional investigation and mitigation plan-
ning. Table 1 contains a summary of all 17 failure incidents and the two premature retirement cases.
Failure incidents have occurred while operating CT with grades 80, 90 and 100 ksi, supplied by three
major US-based CT suppliers, and in size ranges from 1.5-in. conventional to 2.375-in by 1.25-in.
concentric strings.
SPE-179096-MS 5
The incident frequency is very small when compared the cases reported in SPE 81724 (Larsen &
Reichert 2003). That paper details a greater incident frequency (28 cases over 3 years) and includes details
of field failures, an associated laboratory abrasive testing program and a risk categorization system. As
many of the concepts from Larsen & Reichert are expanded on in this paper, their conclusions are
summarized:
Abrasion failures can occur with very low running footage.
Chrome completions, combined with a gas environment and deviated/tortuous well paths, are
particularly hostile.
6 SPE-179096-MS
Abrasion lab testing illustrated the wear rate was negatively affected by dry environment and
increased normal load.
Abrasion lab testing illustrated that material did not have a significant effect on wear rate.
A risk matrix was presented detailing low/medium/high risks for: CT stiffness, maximum dogleg
severity, applied normal force, level of fluid and lubricant presence.
A prediction of running footage was presented with respect to wall thickness for various risk
factors.
Azerbaijan (2014)
A five-well concentric jet pump campaign (new 2.375-in. 1.25-in. 90 ksi material) was conducted
between Dec 2013 and May 2014. The cleanout operation in the first well (Case 11 in Table 1) was
completed in two runs to depths of 11,000-ft and 14,600-ft when scratch marks were noted. For
subsequent operations the wells were fluid filled by adding a 2% lubricant solution in seawater. The
original string was used for two more wells (total of 13 runs and 196,000 running ft) with a third-party
inspection tool used to measure OD, ovality, wall thickness and metal loss/MFL amplitudes. The
inspection did not reveal any significant wall loss, and the shallow depths of the external scratchings could
not be accurately measured. Photographs of the scratches in the third-party report looked very similar to
Fig. 8b below, implying no significant damage increase had occurred subsequent to the original well.
However, given the high-profile and offshore nature of the project, it was decided to prematurely retire
this string. Ensuring a fluid-filled and lubricated completion, two additional wells were completed with
the replacement string. However, on the first operation with the replacement string (Case 13 in Table 1)
there were operational issues in keeping the well liquid filled and lubricated. As a result, similar scratch
marks were noted after three runs. Field practices were reviewed and 5 subsequent runs with 84,000
running ft were completed for the string without further scratching.
SPE-179096-MS 7
Figure 8a & b3D Deviation Plot of Case 11 Well & Observed Scratches
Figure 9 3D Deviation Plot of Well Where Final Damage for Cases 14, 15 & 16 Occurred
Surface burst failure while POOH at a depth of 5,000 ft after two runs in the well
Failure cause was extreme abrasion causing a wall loss of 55% (Fig. 10a)
Figure 10 A, B & CA Left Case 14, B Mid Case 15 & C Right Case 16
Deviation of 60 degrees during step out, increasing to 90 degrees through the formation (see Fig.
11)
Three locations of compressive failure (Fig. 12A) observed on the pipe intrados following 14 runs
and 180,000 rft
Figure 12 A, B & CA Left Compressive Failure, B Additional Waviness, C Right Abrasion Width
Metal shavings, subsequently confirmed as CT material, observed in the stripper on early runs
Subsequently, pipe parted during surface spooling operations (Fig. 2 right)
Root cause of failure external abrasion locally removing 60% of the string wall thickness across
~ 2.6-in. (40% of the circumference) Fig. 12C
Additional pipe waviness observed in the area of the failures (see Fig. 12B)
Failure was not associated with a bias weld
Fluid height in the well was between 500 to 1,000 ft MD (vertical section) for 12 of the 14 runs
well completion. The recovered metal shavings presented a highly plastic deformed ferrite-pearlite
microstructure, similar to the coiled tubing microstructure.
Additional Analysis
Figure 13 A & BCase 15 A Left 3D Well Profile Plot, B Right Dev & DLS v MD
The resulting axial contact forces from pulling from 17,000 ft with 43,000 lbf weight and 58%
combined stress are shown in Fig. 14. The contact force very closely matches the dog leg severity plot.
Please note that contact force is the resultant axial load. The applied normal load would be axial load/0.24
(coefficient of friction used in this case).
The abrasive footage distance was estimated from the string cycles graph as shown in Fig. 15. In this
case the failure location was at 13,700 ft, and 8 passes were made on this well to 15,000 ft, with additional
shorter passes at approximately 14,500 ft and 15,500 ft, yielding a RIH abrasive distance of ~13,200 ft.
12 SPE-179096-MS
The same process as above was applied to the Azerbaijan and US wells, and the results are shown in
Table 2. To provide a comparison, two additional chrome completions were reviewed (a typical Eagle
Ford shale completed with L80 and a Gulf of Mexico deepwater completion Cr13 in both cases abrasion
issues were not observed) and their contact loads, length of deviation included in the table. The associated
3D deviation profile and deviation/DLS v measured depth plots for cases: 11, 13, 18, Eagle Ford and Gulf
of Mexico are shown in the appendix as Figs. 1620.
Table 2Comparison of Abrasive Distance, Dog Leg Severity & Length, Axial Contact Force
POOH Associated
String Fail or damage RIH abrasive Max DLS Max Deviation Length of Build Axial Contact Force
noted or comparison footage (ft) (deg/100 ft) (deg) (ft) (lbf/ft)
The above analysis of damage or failure incidents isolated to single well yields the following:
Focusing exclusively on maximum DLS and maximum deviation as risk factors can be misleading.
The max DLS may occur in the lateral as an azimuth change.
Eagle Ford shale operations have high contact force/high DLS but no history of abrasion
failures.
Additional well profile information is significant.
Duration of the build/DLS is shown to affect failures or initial abrasion evidence.
The New Zealand failure occurred with low: abrasive running footage, DLS, and contact force,
however, a relatively long build section was present.
Force plots showing areas of increased contact friction tend have similar high points as DLS plots,
especially in high step out wells.
Effect of Local Pipe Undulations
The applied contact forces shown in Table 2 above assume that a uniform load distribution is present. As
mentioned earlier, the majority of abrasive failures have some form of local additional bending present.
SPE-179096-MS 13
Localized bending will result in greater local contact forces than the uniform contact loads in Table 2. The
bending occurs from residual curvature close to the BHA, short duration changes in material stiffness (step
bias welds, field butt welds), or other undulations present in the string.
A simplified method of illustrating the increase in magnitude of local normal forces for point contact
can be calculated from the forces required to constrain residually curved pipe in a completion.
The normal forces required to constrain 90 ksi CT of varying sizes are shown in Table 3 below. In each
case the length being constrained was equivalent to 1/8 or 45 degree of a full circle residually curved CT.
Table 3Comparison of Normal Constraining Forces to Displace Residually Curved CT Within a 5.5-in. Casing
Gooseneck Radius Residual Normal Force to
CT Constrained Length Reel Diameter Radius Curvature Constrain
Description (ft) (in.) (in.) (in.) (lbf)
The rise in normal force as pipe stiffness increases illustrates that most mid-string abrasion failures
occur in larger sized pipe. The cases in Table 1 for the 1.5-in. CT are all local end effects and, being close
to the BHA, have exposure to abrasion on the complete wellbore.
Several of the cases in this paper relate to concentric operations (insert case or string numbers when
completed). Assuming that the inner string does not affect the residual curvature, the increase in normal
force for a concentric versus standard string is minimal.
The data in Table 3 assume a uniform pipe and illustrates the order of magnitude of theoretical
constraining forces. By the very nature of local undulations, non-uniform discontinuities (kinks) are
present that will change the above values and can be more accurately modelled using finite element
analysis. The presence of a significant local kink would result in significant contact forces being applied
throughout the completion regardless of DLS.
C Mn Mo Cr
Grade min. max. min. max. min. max. min. max. Ni max. Cu max. P max. S max. Si max.
Grade min. max. Tensile Strength min. (Ksi) Hardness max. (HRC)
L80 80 95 95 23
L80-13Cr 80 95 95 23
P110 110 140 125 -
L80 and L80-13Cr are controlled-hardness grades, and the maximum hardness that is specified in the
standard is 23 HRC. P110 has no limit regarding hardness, but the normal values can be 22 HRC to 29
HRC. This can be compared to the typical CT values of 80/90 ksi 18 20 HRC, 100 ksi ~ 25 HRC and
110 ksi of ~ 30 HRC. Not a significant difference in material hardness between the CT and typical
completions.
API 5CT in the section 7.12 establishes for L80 13Cr that the internal surface of the pipe shall be
free from scale after final heat treatment. In addition, it indicates the possibility for the purchaser to
specify additional requirements regarding the internal surface preparation in Annex H, section H.8. This
section establishes that the pipe inside surfaces shall meet the requirements for Sa 2-1/2 in ISO 8501-1
Preparation of Steel Substrates Before Application of Paints and Related Products - Visual Assessment
of Surface Cleanliness - Part 1: Rust Grades and Preparation Grades of Uncoated Steel Substrates and of
Steel Substrates After Overall Removal of Previous Coatings. Basically, this standard provides repre-
sentative photographic examples of the color changes imparted to steel that is dry-blasted with different
metallic and non-metallic abrasives.
Sa 2-1/2 blast-cleaning finish (near white metal) implies that mill scale, rust paint and foreign matters
are removed completely; any remaining traces are visible only as slight stains or discoloration in the form
of spots or stripes (see examples below Fig. 21).
As it was shown in Fig. 4, the external abrasion damage always presented high external roughness,
inferring that it is caused by the contact with a high roughness surface. From Livescu and Craig (2014)
linear coefficient of friction increases as the surface roughness of the completion increases (Table 6).
From this, it can be established that the total friction force throughout the wellbore, and in consequence
the applied normal load, will increase the severity of abrasion damage as the well completion internal
roughness increases. Any treatment to reduce this internal roughness, like the dry-blasting included as a
Supplementary Requirement in API 5CT for 13 Cr materials, would be helpful in reducing the risk for
external abrasion.
Remediation Plans
Based on the above incidents current remediation plans focus on the following practices.
Risk Identification
Utilize the abrasion risk assessment matrix per Larsen & Reichert (SPE 81724)
Review duration of dog leg severity & contact friction force data
Review CT string make up and avoid using if double step wall thickness changes are present
Obtain surface finishing specifications for chrome completions from well operator
Operational Mitigation
Where practical fluid fill the well
Utilize metal to metal lubricant around upper deviation build areas
Utilize a pipe straightener, if rig up space permits, to remove observed undulations (only
available method to remove the increased damage potential caused by
undulations/deformations)
Observation for damage progression
Incorporate steps in the job program for field crews to pay more attention to pipe condition on
high abrasion risk wells
Incorporate steps in the job program to spot check pipe wall on the job using Ultrasonic
Thickness probe on high abrasion risk wells
Incorporate steps in the job program to check for metal shavings in the stripper / pressure control
equipment between runs on high abrasion risk wells.
Stop and examine all waviness and undulations when RIH
Flag and observe same sections for the onset of abrasion damage
If available, utilize real-time software to flag areas requiring future inspection
Consider use of continuous CT damage monitoring devices
Derating Damaged Section
16 SPE-179096-MS
Further Investigations
Opportunities exist for further investigations. These include:
A better understanding of the cause(s) behind CT undulations/local deformations.
Investigate the additional local loads that these deformations cause
Is CT speed a factor? Does this affect local temperatures that increase the risk of galling? Would
slower running and pulling speeds reduce heat buildup or would faster speeds/reduced time at
temperature mitigate the problem?
Investigate well profiles in Cr13 completions where abrasive damage has not been observed.
Incorporate automated reviews in job design software to highlight and warn of potential abrasion
issues.
Incorporate additional options in job design software to specify completion material and expected
surface roughness.
Conclusions
While statistically not a major CT failure cause, abrasive damage is of concern due to the quick onset of
failure leading to HS&E concerns and operational and economic challenges. In all failure cases, the failure
occurred as excessive wall loss over a short section, with evidence of external roughness and galling
present. In addition, most cases occurred in sections of the CT string where local contact loads were
present due to non-straight pipe in the area of the BHA, CT stiffness changes at tapered bias welds or in
the presence of local deformations. Four cases were examined where the abrasive damage could be
identified to exposure to a single well. Common to these well profiles were long sections in the upper
completion of deviation change not displaying significant dogleg severity. In two cases the build section
was dry, in the other two wells fluid levels were near to surface. Three of the wells were Cr13 completions
with one being L80 (the only mid-string L80 failure noted in the database). Updated remediation practices
and plans for future investigations were created.
Abrasive failures can affect all sizes/grades of CT (even smaller sizes)
Abrasive failure can occur after just 13,200 running ft! Need to be able to identify at the job
planning stage high risk wells
Contact force is a quatitative measure of abrasion risk and closely matches DLS
Can be managed (relatively simply) at an operational level lubricate fluids/keep well full/spot
check pipe/ pipe straightener
Acknowledgements
The authors wish to acknowledge the management of Baker Hughes Inc for permission to publish this
paper. In addition to the authors wish to thank Adrian Terry and Preston Saunders for previous internal
technical studies and Noel Atzmiller for his diligent manuscript proofreading.
References
API Specification 5CT, Specification for Casing and Tubing, 9th Edition, June 2011 / ISO 11960:20011 Petroleum and
Natural Gas IndustriesSteel Pipes for Use as Casing or Tubing for Wells.
Burgos, R., Mattos, R.F., and Bulloch, S., 2007. Delivering Value for Tracking Coiled Tubing Failure Statistics. Presented
at the SPE/ICoTA Conference held in Houston, Texas, 20 21 March 2007. SPE-107098. http://dx.doi.org/10.2118/
107098-MS
SPE-179096-MS 17
Crabtree, A., R., 2008. CT-Failure Monitoring: A Decade of Experience. Presented at the SPE/ICoTA Conference held in
Houston, Texas, 12 April 2008. SPE-113676. http://dx.doi.org/10.2118/113676-MS
ISO 8501-1: 2007 Preparation of Steel Substrates Before Application of Paints and Related Product (Sa 2 ).
Larsen, A.H. and Reichert, B.A., 2003. Coiled Tubing AbrasionAn Experimental Study of Field Failures. Presented at
the SPE/ICoTA Conference held in Houston, Texas, 8 9 April 2003. SPE-81724. http://dx.doi.org/10.2118/ 81724-
MS
Livescu, S., and Craig, S.H., 2014. Increasing Lubricity of Downhole Fluids for Coiled Tubing Operations. SPE Journal
2014. SPE 168298. http://dx.doi.org/10.2118/168298-PA
Van Adrichem, W., P., and Larsen, H., A., 2002. Coiled Tubing Failure Statistics Used to Develop CT Performance
Indicators. SPE Journal 2002. SPE 78808. http://dx.doi.org/10.2118/78808-PA
18 SPE-179096-MS
Appendices
Figure 19 Eagle Ford Comparison 3D Deviation Plot & Dev/DLS v Measured Depth
SPE-179096-MS 19
Figure 20 Gulf of Mexico Comparison 3D Deviation Plot & Dev/DLS v Measured Depth